ABSTRACT Lung cancer is the leading cause of cancer deaths in the USA, with an estimated 158,000 deaths in 2016. The direct cause of the majority of these deaths is the eventual emergence of resistance to initially effective therapies. This evolution of drug resistance represents one of the greatest unmet needs in oncology. While most research is focused on the individual molecular alterations that confer this resistance, we instead propose to focus on the eco-evolutionary processes that generate these alterations. To study the Darwinian evolution and ecological interactions occurring within heterogeneous tumors, we will tightly integrate bespoke mathematical models and experimental techniques designed to inform them. Focusing on EGFR mutated non-small cell lung cancer, a cancer type with a highly efficacious targeted therapy with which we have experience in our lab, we will approach this problem with three, orthogonal, integrated mathematical-experimental Aims. First, to understand the ecological interactions occurring at the inter-cellular level in heterogeneous tumors, we will couple our experience with evolutionary game theory with our first-in-class evolutionary game assay, which we have designed to specifically for this purpose. Here, we hope to learn to target the interactions that drive resistance ? a novel strategy which could open up entirely new avenues of drug design. Second, we will allow evolution to show us the convergent phenotypes it creates in the face of specific selective pressures through long-term directed evolution. During this long-term evolution, we will measure phenotype, in the form of drug sensitivity to a panel of chemotherapeutics and targeted therapies at regular intervals, creating the first temporal collateral sensitivity map in any solid tumor. By pooling common phenotypes observed throughout the evolutionary life history, we will then use interactomic and seed-based protocols to generate molecular signatures of these states of sensitivity, which we will validate in publicly available data and in an in library of PDX lines. Finally, we will delve deeply into the relevant time scales of the ecological and evolutionary processes we study in the first two aims. We plan to apply the replicator-mutator framework of evolutionary game theory to a spatial transform that we pioneered in cancer. To validate and parameterize these models, we will also test the evolutionary stability of the ecological dynamics we measure with our game assay by performing the assay through evolutionary time during a long-term evolution experimental. Each of our three orthogonal aims is supported by recent high impact publications, and each represent tightly coupled experimental and computational protocols. Our clean, well- designed integration, together with innovative focus on the direct study of the evolutionary biology, promises to shed light on this difficult area of cancer research, and offers the possibility of providing generalizable insights.